Volcano Watch: What sounds the (automated) alarms at HVO?
by US Geological Survey Hawaiian Volcano Observatory scientists and affiliates
The US Geological Survey Hawaiian Volcano Observatory keeps its eyes on active volcanoes in many ways.
While old-fashioned eyes and a notebook are used when field teams are near a volcano, modern volcano observatories also use rapidly collected data and computers to support monitoring.
Because it can be very expensive to have scientists watch volcano data streams on a 24-hour basis, Hawaiian Volcano Observatory computers are “trained” to look for activity and send alerts when a volcano is changing or becoming active.
The computers look at a wide array of observations including seismic, or ground shaking; infrasound data, or air pressure; and ground deformation, as well as camera image assessments.
Data from these systems can be assessed as close as possible to the time of collection as it enters a computer. That computer can send a message to alert a scientist at any time of day or night if an observation is unusual.
At the core of many alarm systems is an often-simple computer program that looks for a change in energy from shaking of a seismometer or a burst of sound on an acoustic sensor. If the sensor is usually quiet and then a burst of energy occurs, it can be detected by the computer which can send an automated message to a scientist.
This type of energy burst detector is called a short-term average/long-term average detector.
Changes in the air pressure during an eruption are recorded on sound detecting sensors, such as like microphones, called an infrasound or acoustic array.
The figure of four graphs included with this week’s “Volcano Watch” shows how a successful detection of an eruption might work using infrasound data and how the system could fail to detect an eruption.
The first image (A) shows an eruption that creates pressure changes recorded by the sensor.
Detection, marked by a red arrow, would be easy to determine using a short-term average/long-term average method, discussed above and shown as the gray and red bar at the bottom of panel A. The long-term average, or gray bar, is a period before the eruption and the short-term average, or red bar, shows when the eruption energy is strong.
In this case, the short-term average is much bigger than the long-term average and we can set the computer to send a message when this specific condition happens.
We then consider the same sensor that detected another burst (B), but this burst came from a non-eruption source — say a car, helicopter or another event away from our volcano. In this case, the volcano scientist might be alerted by an event that wasn’t volcanic, or a false detection.
In the next example we introduce wind noise (C) into the same eruption recorded in (A). In this case, the wind would be so strong that the eruption can only just be seen in the data by the naked eye but might not be visible to the computer’s short-term average/long-term average detector.
The last example adds wind noise that might have been a car or helicopter but not an eruption (D). Here, the non-eruption can only barely be seen in the data and was not detected by the short-term average/long-term average detector.
The figures show four possible outcomes that can be used to assess how good an alarming system performs.
They include:
- Detection of a real eruption (A).
- A false detection of non-volcanic change (B).
- Failure to detect a real eruption (C).
- Not detecting any event of interest (D).
These four examples show how Hawaiian Volcano Observatory scientists can assess the performance of our alarm systems to improve detection of volcanic events and minimize the detection of noise or other activity.
The four examples coincide with the concepts of test conditions for alarms using the terms true-positive (A), false-positive (B), false-negative (C) and true-negative (D).
A good alarm system should include true-positives (volcanic events) and true negatives (non-volcanic events) while trying to minimize false-positives and false-negatives. Too many false alerts equate to unnecessary loss of sleep during quiet periods for a volcano.
In summary: alarm systems are an important and evolving part of Hawaiian Volcano Observatory operations and an example of how computers and technology are incorporated in monitoring.
Volcano Activity Updates
Kīlauea is not erupting. Its US Geological Survey Volcano Alert Level is Advisory.
Earthquake rates during the past week beneath Kīlauea’s summit and upper East Rift Zone were more than double that of the previous week. About 100 earthquakes were located beneath the summit and about 226 in the upper East Rift Zone.
Earthquake rates beneath the middle East Rift Zone were on par with the previous week.
Ground deformation rates in the summit region showed increased inflation during the past week, while ground deformation rates near the Sept. 15-20 middle East Rift Zone eruption site have slowed.
Future intrusive episodes and eruptions could occur with continued magma supply.
Maunaloa is not erupting. Its US Geological Survey Volcano Alert Level is at Normal.
Three earthquakes were reported felt during the past week in the Hawaiian Islands:
- A magnitude-3.1 earthquake 3 miles south-southwest of Volcano at a depth of 1 mile Nov. 27 at 2:19 p.m..
- A magnitude-3.5 earthquake 46 miles east-northeast of of Honomū at a depth of 20 miles Nov. 27 at 12:29 a.m.
- A magnitude-3.1 earthquake 1 mile south-southwest of Pāhala at at depth of 18 miles Nov. 22 at 2:02 p.m.
Hawaiian Volcano Observatory continues to closely monitor Kīlauea and Maunaloa.
Visit the volcano observatory’s website for past “Volcano Watch” articles, Kīlauea and Maunaloa updates, volcano photos, maps, recent earthquake information and more. Email questions to askHVO@usgs.gov.
Sponsored Content
